To understand the molecules important to our projects on the pharmacology of biological ligands (Z01 CB 08366-07) and combination therapy (Z01 CB 08392-03), we have explored several innovative ways of analyzing protein structure-function relationships: 1. HAL, HALP, & HALCO. We developed these statistical mechanics-based algorithms to evaluate amphipathic helical structures and more general structural motifs in proteins and peptides, including those of the HIV-1 envelope. 2. A joint prediction algorithm. This new method predicts peptide secondary structure by concatenating the predictions of several different algorithms (e.g., based on neural networks, information theory, sequence homology, hydrophobicity, loop potential, and amphipathicity.) This initially ad hoc approach was formalized in an algorithm called Q7- JASEP(for 07-based Joint Algorithm for Secondary Structure Prediction). 3. 3-dimensional quantitative structure-activity relationship (3D-QSAR) studies. 3D-QSAR analysis was done using a set of 20 nucleoside transport inhibitors. The nucleoside transport protein is a major target for dipyridamole, which we have found to potentiate the activity of AZT/dipyridamole in HIV-infected patients are in progress. and the 3D- QSAR predictions can be used to direct the design of analogue inhibitors for study. 4. Thermodynamic cycle perturbation (TCP) on HIV-1 protease-inhibitor complexes. We are using TCP methods to "mutate" one peptidomimetic inhibitor of HIV-1 protease computationally into another and predict the binding free energy of the latter peptide from that of the former. Interest of this analysis arises from the fact that inhibitors of HIV-1 protease are among the most promising new agents for treatment of HIV infection. We are currently collaborating in the analysis of their behavior in combination with dideoxynucleosides (part of Z01 CB 08392- 03).